While there may not be an electrical effect, per se, there will be an effect, however tiny. Many of the responses seem to focus on the tininess of the effect and conflate "really really small" with "doesn't exist".
Specifically (and maybe there are others that I'm not thinking of), the temperature of the sun is a function of the radiation leaving it minus the radiation it receives from around it. This fact is encapsulated in the radiative heat transfer law where heat exchange is equal to the Stefan-Boltzmann constant times the difference of the 4th powers of the temperatures. Because the earth is absorbing radiation from the sun, its temperature increases, increasing the amount of thermal radiation it, in turn, radiates outward. Some very tiny fraction of that light will hit the sun, increasing its temperature by such a tiny bit. To be sure, the effect is compounded by the fact that the earth also reflects some light to the sun, which has the same effect, it just doesn't follow the 4th power law.
In other words, the earth is a tiny piece of insulation, blocking the heat of the sun from escaping and warming it up ever so slightly.
So when you throw a solar panel in the mix, you're actually (and usually very temporarily) converting some of that radiation into ordered work instead of pure heat, So the temperature of the earth increases just slightly less, causing just slightly less heat radiation back to the sun.
The are other effects of course; since solar panels absorb most light and reflect very little, they may actually increase the absorption of heat from the sun, although the albedo of black asphalt single is also pretty low (which is why the white roofs are mentioned elsewhere). So if the albedo of the solar panels plus their efficiency in converting light to electricity is less than the albedo of the background (the roof), then the temperature of the earth actually goes up, but the reflection goes down. Either way, there's an effect.
I think part of the confusion lies in the fact that many (maybe all? ) of our physical laws, especially those dealing with wavelike phenomena are pragmatic approximations. So things like the near-field effect don't suddenly vanish after a few wavelengths, they just become so small that they make no difference in the predictive power of those models for practical use. Which is to say, they become smaller than other sources of error in the prediction of how the system will behave. Negligible does not mean nonexistent.